Download Perception of English Vowel Reduction by Trained Spanish Learners

Perception of English Vowel Reduction by Trained Spanish Learners
Esther Gómez Lacabex1, María Luisa García Lecumberri1, Martin Cooke2
1
2
University of the Basque Country, Spain
University of Sheffield, United Kingdom
[email protected], [email protected]
1. Introduction
It has been widely acknowledged that speech perception of non-native contrasts in second
language acquisition (SLA) poses difficulties for L2 speakers (Best, 2001). Factors such as
the influence of the L1 phonological system, markedness (Eckman, 1977) universal
tendencies or individual characteristics may account for some of these difficulties.
Nevertheless there are studies (Bongaerts et al., 1997; Markham, 1997; Escudero, 2006) that
have shown that L2 speakers can learn to perceive new L2 vowels in a native-like manner and
theories like the Speech Learning Model (SLM) by Flege defend the idea that the native
language sound system mechanisms ‘remain intact over the life span’ (Flege, 1995:239) and
that it is possible for adult learners to establish new non-native phonetic categories.
In line with the idea that it is possible to acquire a new sound category, our study aims
at assessing the ability to perceive the English unrounded mid-central vowel schwa by
Spanish learners of English. While schwa is a very frequent sound in English as the nucleus in
unstressed syllables both in content words and function words, the Spanish vowel space has
an empty central area with no central vowel categories. This fact in addition to others such as
the influence of orthography, specially in contexts in which language is introduced in its
written form (most formal instruction settings) may explain why this sound tends to be
assimilated to L1 vowel sounds by Spanish speakers (and therefore identified and produced as
a peripheral vowel).
Different views can be found in the literature on vowel reduction. It has been
explained as a correlate of stress when described as the shortening in duration and obscuration
of quality of nuclear vowels in unstressed syllables (Rietveld & Koopmans-van Beinum,
1987). It has also been defined as ‘formant undershoot’ or the failure of formant frequencies
to reach target representations when duration is shortened (Lindblom, 1963). While this type
of vowel reduction is very frequent in the languages of the world, the reduction degree
reached differs between languages. As afore mentioned, vowel reduction in Spanish (Delattre,
1969) involves slight centring movements only and vowels never lose their quality (Quilis,
1996). In English, however, reduced vowels have become part of its phonological inventory,
that is, they have become phonemes or sound categories, and thus, distinctive in contrasts of
the type full vowel–weak vowel in unstressed syllable (exorcise- exercise).
Previous studies on the acquisition of vowel reduction in SLA have shown that stress
and vowel reduction are treated as independent phenomena by Spanish speakers (Flege &
Bohn, 1989) or that learners show some awareness of schwa but fail to produce contextual
coarticulatory patterns (Kondo, 1995). We also find studies on the effect of treatment on the
acquisition of vowel reduction. Gutierrez and Monroy’s study (2003) tested a similar age
group (15-17) with the same L1 (Spanish) as our study on the same type of vowel reduction
(‘phonological’ as in ‘private’ or ‘suspect’), and concluded that there was no improvement on
the production of vowel reduction after the training. While the authors refer to the fact that the
training period was too short in order to account for the lack of treatment effect, these results
could be due to other factors such as type of training administered. A further study which
analysed the perception of the contrast full vowel-reduced vowel by Spanish learner of
English without specific training revealed that perception of vowel reduction only occurred at
chance performance levels (56% correct answers) (Gómez Lacabex & García Lecumberri,
2005).
This study also aimed at examining the relationship between perception and
production training. Auditory and articulatory treatments were administered to two different
training groups in order to explore potential different outcomes between them. Diverse
findings may be found in the literature as far as this relationship is concerned. Studies like
those of Lively et al. (1994) or McClasky et al. (1993) reported positive effects in laboratory
discrimination training on perception abilities while positive effects of production training on
production abilities have also been reported (Flege, 1988). Interestingly, a close relationship
between production and perception is supported by those studies that have found that training
in one of the skills exerts positive influence on the other. Leather (1990) analysed the effect of
training on Chinese lexical tones on production and perception: while one group received
computer-based training in perception, the other group was trained to use computer-managed
visual feedback on their production. The groups were tested on both perceptual and
production abilities and results suggested that both training types had positive effects since
training in one modality ‘tended to be sufficient to enable a learner to perform in the other’
(Leather, 1990: 95).
In sound training research, generalization has often been used as a measure of learning
robustness (Lively et al., 1994; Hardison, 2004) Our study also aimed at investigating whether
the two types of training provided (perceptual and articulatory) were equally effective as far
as robustness is concerned. Studies like those of Rochet (1995) or, McClasky et al. (1983)
have concluded that generalization after a training period can occur. More precisely, Rochet
found that his adult subjects (native speakers of Madarin Chinese) were able to transfer voice
onset time (VOT) values across consonant place of articulation and vocalic contexts after a
perceptual training with natural speech tokens. McClasky et. al reported similar findings, this
time with synthetic stimuli, as their subjects (monolingual speakers of English) showed
significant gains after discrimination laboratory training on voice onset time values in place of
articulation which had not been used in the treatment sessions.
2. Method
2.1 Participants and procedure
This study selected a group of 50 subjects to take part in the experiment (two experimental
groups and a control group). However, all groups suffered attrition, the control group being
most severely affected (from 12 subjects in pre-test to 7 subjects in post-test). Eventually, data
were obtained from 41 Spanish teenagers (24 female, 17 male, mean age: 15.8) learning
English as a FL in a formal learning context. They attended English lessons for 3 hours a
week at a private language school -where the training took place- in addition to English
instruction at school. Subjects had been studying at the private language school for 4.1 years
on average. This was an important factor as it involved exposure to native English accents,
which they did not have in the state school. Subjects were divided into three groups: two
experimental groups (A and B) undergoing two different types of vowel reduction training
(perceptual or productive) and a control group (C), which was not given any specific training
(see Table 1 for subject and gender distribution among groups). The training period was
conducted over 3 months. Students received specific training on vowel reduction once a week
(12 sessions on average) within their English course. Perceptual training was based on
discriminatory exercises and production on the part of the students was not encouraged.
_____________________
New Sounds 2007: Proceedings of the Fifth International Symposium on the Acquisition of Second Language Speech
- 294 -
Production training, on the other hand, provided students with articulatory and visual cues and
was based on production of the items on the part of the students and individual feedback
provided by the trainer. Here, perceptual activities were avoided and sample/model
productions were limited in the classroom.
Table 1. Distribution of subjects and gender for the three groups
N
gender
male female
ABC 41 17
24
A
17 10
7
12
B
17 5
C
7 2
5
2.2 Stimuli
A total number of 22 items were presented to the subjects in a two alternative forced choice
identification task (AX) with orthographic input which was custom designed in MATLB. This
corpus consisted of 11 minimal word pairs with the contrast ‘full vowel’ versus ‘schwa’1 in
the unstressed syllable (i.e., seafood-Seaford). Prior to testing, students were familiarized with
the stimuli so that they would know their meaning. These minimal word pair sets were also
presented in the carrier sentence: I’m going to say … again. Thus, vowel perception was
presented at two different levels: word level and sentence level. Here we present results
corresponding to the perception of the weak words (11), that is, those words with schwa as the
nucleus of the unstressed syllable.
minimal word pairs
weak
(analysed in this study)
turban
datum
Normanton
commanders
Seaford
salad
pillars
hearers
Seaward
Upton
starlet
strong
turbine
daytime
Normantown
commandoes
seafood
sallowed
pillows
heroes
seawood
upturn
starlight
3. Data analysis and results
Students' perceptions were coded as percentages of correct answers. A repeated measures
ANOVA of the results of the combined experimental groups (A&B) versus the control group
(C) in the pre and post training conditions was carried out (Figure 1) and a further analysis
1
Due to the difficulty of finding short vowel vs. schwa minimal pairs, the contrasts presented here were of the
type long vowel-schwa. Being aware of this limitation an additional corpus -which is not analyzed in the present
study- was created which contrasted short-full vowels vs. schwa as in Superman-fireman at syllable level.
_____________________
New Sounds 2007: Proceedings of the Fifth International Symposium on the Acquisition of Second Language Speech
- 295 -
pre-test
posttest
100
90
80
70
60
50
40
A&B
identification score %
identification score %
compared the experimental group receiving perception training (A) with the experimental
group receiving production training (B) (see Figure 2).
Inter-group analysis showed no significant differences between the experimental
groups and the control group in the pre-test neither in words or sentences. Near significant
differences were found between A&B and C in the post test for words (F(1,39) = 2.91,
p=.096) and sentences (F(1,39) = 3.81, p=.058). They probably did not reach significance
due to the large standard deviations in group C.
Intra-group analysis for experimental groups A&B taken together revealed significant
differences between pre-test and post-test at word level perception (F(1,39) = 15.78, p<.0001)
and at sentence level perception (F(1,39) = 11.44, p<.005). For control group C pre-test and
post-test comparisons showed no significant differences at word nor at sentence level.
A further inter-task analysis revealed that there were no differences between
perception at the word level and perception at the sentence level. This was found to be
consistent across groups and treatment stages (p>.05).
pre-test
posttest
100
90
80
70
60
50
40
C
A&B
word
C
sentence
Figure 1. Perception of vowel reduction in words (left) and sentences (right) contrasting experimental
groups together (A&B) and control group (C)
When comparing how the two experimental groups performed individually, intergroup analysis showed no differences between experimental group A and experimental group
B in any of the perception tasks either in the pre-test (word level: F(1,32) =.32 p=.576 and
sentence level: F(1,32) =.72 p=.403) or in the post-test (word level: F(1,32) =.02 p=.888,
sentence level: F(1,32) =.00 p=1.00).
Intra-group analysis revealed significant/near significant differences between pre-test
and post-test in both contexts and for both groups. At word level, both groups showed
significant improvement (A: F(1,39) = 6.72, p<.05; B: F(1,39) = 14.64, p<.005); for
perception of schwa at sentence level, it was observed that while improvement on the part of
group B was significant, improvement in group A did not reach statistical significance (A:
F(1,39) = 3.21, p=.083; B: F(1,39) = 8.91, p=.005). In order to explore group differences,
correlation measures between students' perception scores and gender were computed; no
significant correlations were found.
Further inter-task analysis revealed that there were no differences between perception
at the word level and perception at the sentence level between experimental groups A and B
either in pre-tests condition or in post-test condition (p>.05).
_____________________
New Sounds 2007: Proceedings of the Fifth International Symposium on the Acquisition of Second Language Speech
- 296 -
90
80
70
60
50
40
A
identification score %
identification score %
pre-test
posttest
100
pre-test
posttest
100
90
80
70
60
50
40
A
B
B
sentence
word
Figure 2. Perception of vowel reduction in words (left) and sentences (right)
contrasting experimental groups separately
4. Discussion
In light of the data obtained, we can say that our students were able to improve their
perception of schwa at the word level after the specific training administered; that is, the
training proved effective since both experimental groups displayed an improvement as
differences between testing stages reflect. Moreover, the control group did not show
improvement indicating that the sole exposure and language experience gained in the period
between pre and post tests was not responsible for the improvement displayed by the
experimental groups. There were near significant differences between the experimental
groups and the control group in the post-test for perception of schwa at word level. This lack
of significance was probably due to group C’s attrition and large standard deviation, so a
larger control group would be necessary to explore this tendency further. Results agree with
previous studies discussed above which have shown that non-native sounds can be learnt and
support theoretical views which contemplate the possibility of developing perceptual abilities
for non-native categories, in particular for sounds which may be perceived as ‘new’ due to
their lack of similarity to L1 categories (Flege, 1995; Best, 2001) as is the case of schwa for
Spanish speakers.
Generalization of perception of schwa to a non-instructed and new context (sentences)
occurred in our experiment. Students were not trained in perceiving schwa within a sentence,
a more complex and more natural context, nevertheless they showed significant improvement
in their perceptual abilities in this context after the treatment (Figure 1 right). While the
control group did not experience any improvement, near significant differences were found
between A&B and the control group in the post test, supporting the treatment effect. In
addition, this generalization happened in both experimental groups as Figure 4 shows. Our
results agree with other studies which have shown the ability to generalize to non-instructed
contexts (Rochet, 1995; McClasky et al., 1983). This result is only a first stage towards
exploring learning robustness since other types of generalization (to non-instructed words and
other speakers) are being studied.
Results revealed no significant differences between experimental groups A and B, that
is, the two treatments administered in our study did not produce significantly different results.
Given the fact that experimental group A received instruction based on perception, the type of
task and the type of skills worked on during the treatment were similar to those used in the
pre-test and post-test. Group B, on the other hand, received a different training approach
based on articulatory cues and production feedback. This being so, group A would have been
expected to perform better that group B as students were trained in the task demanded in the
testing sessions. However, results indicate that both groups performed similarly in the
_____________________
New Sounds 2007: Proceedings of the Fifth International Symposium on the Acquisition of Second Language Speech
- 297 -
perception test. Our results suggest that production training exerts a positive influence on
perceptual abilities being as efficient as perceptual training. At this point our study supports
the facilitating view between perception and production training reported in Section 1 above
in one direction: production training facilitating perception. Further analysis of the students’
productions will reveal whether this view is fully supported.
5. Conclusions
Our study has shown that specific training on English vowel reduction in lexical words
produced a positive effect on Spanish learners’ perceptual skills. Results support recent
theoretical views regarding the possibility of developing perceptual abilities when identifying
non-native sounds and developing new phonetic categories.
Our results did not show differences in perception performance between auditory and
articulatory training regimes. The group receiving auditory training could have been expected
to show some degree of advantage since they were trained in the tests task but this was not so.
We found that articulatory training contributed to the development of the students’ perceptual
abilities, supporting the facilitating relationship between perception and production. Further
analysis of the students’ productions of vowel reduction will answer the remaining question
of whether perceptual training actively contributes to developing students’ production
abilities.
The present study also revealed that the improvement in perception carried over to a
non-instructed more challenging context. This type of generalization is a first indication of
robustness which will be further explored for schwa perception in novel or non-instructed
words.
References
Best, C., McRoberts, G., & Goodell, E. (2001). Discrimination of non-native consonant contrasts
varying in perceptual assimilation to the listener’s native phonological system. Journal of the
Acoustical Society of America, 109(2), 775-794.
Bongaerts, T., van Summeren, C., Planken, B., & Schils, E. (1997). Age and ultimate attainment in the
pronunciation of a foreign language. Studies in Second Language Acquisition, 19, 447-465.
Delattre, P. (1969). An acoustic and articulatory study of vowel reduction in four languages.
International Review of Applied Linguistics, 6(4), 295-323.
Escudero, P. (2006). The phonological and phonetic development of new vowel contrasts in Spanish
learners of English. In B. O. Baptista, & M. A. Watkins (Eds.), English with a Latin beat: Studies
in Portuguese/Spanish English interphonology. Amsterdam: John Benjamins.
Eckman, F. R. (1977). Markedness and the Contrastive Analysis Hypothesis. Language Learning, 27,
315-330.
Flege, J. E. (1988). Using visual information to train foreign-language vowel production. Language
Learning, 38(3), 364-407.
Flege, J. E., & Bohn, O.-S. (1989). An instrumental study of vowel reduction and stress placement in
Spanish accented English. Studies in Second Language Acquisition, 11, 35-62.
Flege, J. E. (1995). Second-language speech learning: Theory, findings and problems. In W. Strange
(Ed.), Speech perception and linguistic experience: Theoretical and methodological issues (pp.
233-277). Timonium, MD: York Press.
Gómez Lacabex, E., & García Lecumberri, M. L. (2005). English vowel reduction by untrained spanish
learners: Perception and production. PTLC 2005, http://www.phon.ucl.ac.uk/ptlc2005.html/.
Gutierrez, F., & Monroy, R. (2003). Proceedings of the 3rd International Contrastive Linguistics
Conference, September 2003, 211-222.
_____________________
New Sounds 2007: Proceedings of the Fifth International Symposium on the Acquisition of Second Language Speech
- 298 -
Hardison, D. M. (2004). Generalization of computed assisted prosody training: Quantitative and
qualitative findings. Language Learning and Technology, 8(1), 34-52.
Kondo, Y. (1994). Targetless schwa: Is that how we get the impression of stress timing in English?
Proceedings of the Edinburgh Department Conference ’94, 63-76.
Leather, J. (1990). Perceptual and productive learning of Chinese lexical tone by Dutch and English
speakers. In J. Leather, & A. James (Eds.), New Sounds 90: Proceedings of the Amsterdam
Symposium on the Acquisition of Second Language Speech (pp.305-341). Amsterdam: University
of Amsterdam.
Lindblom, B. (1963). A spectrographic study of vowel reduction. Journal of the Acoustical Society of
America, 35, 1773-1781.
Liveli, S. E., Pisoni, D. B, Yamada, R. A, Tohkura, Y, & Yamada, T. (1994). Training Japanese
listeners to identify English /r/ and /l/. III Long-term retention of new phonetic categories. Journal
of the Acoustical Society of America, 96(4), 2076-2087.
McClasky, C. L., Pisoni, D., & Carrell, T. D. (1983). Transfer of training of a new linguistic contrast
in voicing. Perception and Psycholinguistics, 34(4), 323-330.
Markham, D. (1997). Phonetic imitation, accent and the learner. Lund University Press.
Quilis, A., & Fernández, J. (1996). Curso de fonética y fonología españolas para estudiantes
angloamericanos (15º Edición). Madrid: Consejo Superior de Investigaciones Científicas.
Rietveld, A., & Koopmans-van Beinum, C. M. (1987). Vowel reduction and stress. Speech
Communication, 6, 217-229.
Rochet, B. L. (1995). Perception and production of second language speech by adults. In W. Strange
(Ed.), Speech perception and linguistic experience: Theoretical and methodological issues in
cross-language speech research (pp. 379-410). Timonium, MD: York Press.
_____________________
New Sounds 2007: Proceedings of the Fifth International Symposium on the Acquisition of Second Language Speech
- 299 -